3D Protein Models Research Articles

RT-PCR Based Cloning and Sequencing of Potato Leaf Roll Virus-Coat Protein (PLRV-CP) Gene for Characterization as a Bangladeshi PLRV Isolate and Its Phylogenetic Analysis

An experiment was conducted in Molecular Biology and Plant Virology Laboratory under the Department of Plant Pathology, Sher-e-Bangla Agricultural University, Dhaka-1207, Bangladesh. Total RNA was extracted from Potato leaf roll virus (PLRV) positive leaves and complementary DNA (cDNA) were synthesized from total RNA. Reverse transcriptase polymerase chain reaction (RT-PCR) based detection conditions were optimized by using coat protein (CP) gene specific primers. In PCR amplification cDNA and in nucleotide sequencing PCR product was used as a template. A 346 bp amplicon of PLRV- CP gene was amplified and amplified gene region was sequenced. The expected nucleotide sequence of amplified PLRV-CP gene showed 95 to 98% homology when compared with the isolates sequences reported in Gene Bank database. This explored novel PLRV-CP gene was characterized as a PLRV Bangladeshi isolate (Accession number, Bankit 2274496, MN605963). PLRV-CP gene protein modeling was carried out using Expert Protein Analysis System (ExPaSy), DNASTAR’s protein tools server used for 3D protein modeling. Phylogenetic analysis was also carried out, the tree was made by using MEGA 4.0 software and maximum parsimony method was selected to construct phylogenetic tree. The RT-PCR based molecular technique optimized in this study, would be a useful for early detection, epidemiological studies of PLRV as well as in seed tubers certification program and the novel hyper variable sequenced region of PLRV-CP gene will be useful in pathogen derived resistance breeding program against the PLRV local strain.

Cover Caption

June Online Cover: 3D structure models of proteins from Hericium erinaceusfrom and their effects on gut microbiota, from “Auxiliary antitumor effects of fungal proteins from Hericium erinaceus by target on the gut microbiota” by Wang Dongdong, Zhu Xiangxiang, Tang Xiaocui, Li Hongye, Xie Yizhen, and Chen Diling. p. 1872.

Deep Learning Based Drug Screening for Novel Coronavirus 2019-nCov.

A novel coronavirus, called 2019-nCoV, was recently found in Wuhan, Hubei Province of China, and now is spreading across China and other parts of the world. Although there are some drugs to treat 2019-nCoV, there is no proper scientific evidence about its activity on the virus. It is of high significance to develop a drug that can combat the virus effectively to save valuable human lives. It usually takes a much longer time to develop a drug using traditional methods. For 2019-nCoV, it is now better to rely on some alternative methods such as deep learning to develop drugs that can combat such a disease effectively since 2019-nCoV is highly homologous to SARS-CoV. In the present work, we first collected virus RNA sequences of 18 patients reported to have 2019-nCoV from the public domain database, translated the RNA into protein sequences, and performed multiple sequence alignment. After a careful literature survey and sequence analysis, 3C-like protease is considered to be a major therapeutic target and we built a protein 3D model of 3C-like protease using homology modeling. Relying on the structural model, we used a pipeline to perform large scale virtual screening by using a deep learning based method to accurately rank/identify protein–ligand interacting pairs developed recently in our group. Our model identified potential drugs for 2019-nCoV 3C-like protease by performing drug screening against four chemical compound databases (Chimdiv, Targetmol-Approved_Drug_Library, Targetmol-Natural_Compound_Library, and Targetmol-Bioactive_Compound_Library) and a database of tripeptides. Through this paper, we provided the list of possible chemical ligands (Meglumine, Vidarabine, Adenosine, d-Sorbitol, d-Mannitol, Sodium_gluconate, Ganciclovir and Chlorobutanol) and peptide drugs (combination of isoleucine, lysine and proline) from the databases to guide the experimental scientists and validate the molecules which can combat the virus in a shorter time.Electronic supplementary materialThe online version of this article (10.1007/s12539-020-00376-6) contains supplementary material, which is available to authorized users.

Study of the structure and binding site features of FaEXPA2, an α-expansin protein involved in strawberry fruit softening

Tissue softening accompanies the ripening of many fruits and initiates the processes of irreversible deterioration. Expansins are plant cell wall proteins that have been proposed to disrupt hydrogen bonds within the cell wall polymer matrix. Several authors have shown that FaEXPA2 is a key gene that shows an increased expression level during ripening and softening of the strawberry fruit. For this reason, FaEXPA2 is frequently used as a molecular marker of softening in strawberry fruit, and changes in its relative expression have been related to changes in fruit firmness. In this context, we previously reported that FaEXPA2 has a high accumulation rate during fruit ripening in four different strawberry cultivars; however, the molecular mechanism of FaEXPA2 or expansins in general is not yet clear. Herein, a 3D model of the FaEXPA2 protein was built by comparative modeling to understand how FaEXPA2 interacts with different cell wall components at the molecular level. First, the structure was shown to display two domains characteristic of the other expansins that were previously described. The protein-ligand interaction was evaluated by molecular dynamic (MD) simulation using four different long ligands (a cellulose fiber, two of the more important xyloglucan (XG) fibers found in strawberry (XXXG and XXFG type), and a pectin (homogalacturonic acid type)). The results showed that FaEXPA2 formed a more stable complex with cellulose than other ligands via the different residues present in the open groove surface of its two domains, while FaEXPA2 did not interact with the pectin ligand.

QUALITY-MODEL CLUSTERING TOOL: A MODULE FOR CLUSTERING PROTEIN MODELS BASED ON QUALITY ATTRIBUTES

The process of protein modeling usually involves the production of a variety of structures requiring efficient tools for structure model comparison attempting to choose the best three-dimensional (3D) structure. This paper introduces an alternative method for clustering 3D protein models that, instead of using attributes related to structural alignment to group the data, use quality-attributes of those models to represent and cluster them. This method stands out by removing the need to define a priori a base model for structural alignments. Even so, it is possible to present the most representative structure in each cluster, which is useful for docking or molecular dynamics studies. All the results were statistically analyzed and compared with decisions made by professionals to validate the proposed algorithm. The experiment simulated a usual protein comparative modeling process for different CATH classifications. The calculated variance levels after the dimensional reduction validate the workflow for different protein chain sizes. All the molecular descriptors for the input files are calculated by MHOLline 2.0, an online scientific workflow for studies on bioinformatics and computational biology, available for free on www.mholline2.lncc.br, or hand made using specific programs (e.g., MODELLER, PROCHECK) and adjusting the data to the template specified in this document. The Quality-Model Clustering Tool (QMC) and the data set used in this work are available for download on the git repository github.com/ruanmedina/Quality-Model-Clustering.

First comprehensive TSC1/TSC2 mutational analysis in Mexican patients with Tuberous Sclerosis Complex reveals numerous novel pathogenic variants

The aim of this study was to improve knowledge of the mutational spectrum causing tuberous sclerosis complex (TSC) in a sample of Mexican patients, given the limited information available regarding this disease in Mexico and Latin America. Four different molecular techniques were implemented to identify from single nucleotide variants to large rearrangements in the TSC1 and TSC2 genes of 66 unrelated Mexican-descent patients that clinically fulfilled the criteria for a definitive TSC diagnosis. The mutation detection rate was 94%, TSC2 pathogenic variants (PV) prevailed over TSC1 PV (77% vs. 23%) and a recurrent mutation site (hotspot) was observed in TSC1 exon 15. Interestingly, 40% of the identified mutations had not been previously reported. The wide range of novels PV made it difficult to establish any genotype-phenotype correlation, but most of the PV conditioned neurological involvement (intellectual disability and epilepsy). Our 3D protein modeling of two variants classified as likely pathogenic demonstrated that they could alter the structure and function of the hamartin (TSC1) or tuberin (TSC2) proteins. Molecular analyses of parents and first-degree affected family members of the index cases enabled us to distinguish familial (18%) from sporadic (82%) cases and to identify one case of apparent gonadal mosaicism.

Whole exome sequencing of a Saudi family and systems biology analysis identifies CPED1 as a putative causative gene to Celiac Disease.

Celiac disease (CD) is a gastrointestinal disorder whose genetic basis is not fully understood. Therefore, we studied a Saudi family with two CD affected siblings to discover the causal genetic defect. Through whole exome sequencing (WES), we identified that both siblings have inherited an extremely rare and deleterious CPED1 genetic variant (c.241 A > G; p.Thr81Ala) segregating as autosomal recessive mutation, suggesting its putative causal role in the CD. Saudi population specific minor allele frequency (MAF) analysis has confirmed its extremely rare prevalence in homozygous condition (MAF is 0.0004). The Sanger sequencing analysis confirmed the absence of this homozygous variant in 100 sporadic Saudi CD cases. Genotype-Tissue Expression (GTEx) data has revealed that CPED1 is abundantly expressed in gastrointestinal mucosa. By using a combination of systems biology approaches like protein 3D modeling, stability analysis and nucleotide sequence conservation analysis, we have further established that this variant is deleterious to the structural and functional aspects of CPED1 protein. To the best of our knowledge, this variant has not been previously reported in CD or any other gastrointestinal disease. The cell culture and animal model studies could provide further insight into the exact role of CPED1 p.Thr81Ala variant in the pathophysiology of CD. In conclusion, by using WES and systems biology analysis, present study for the first-time reports CPED1 as a potential causative gene for CD in a Saudi family with potential implications to both disease diagnosis and genetic counseling.

Model Reconstruction from Small-Angle X-Ray Scattering Data Using Deep Learning Methods.

SummarySmall-angle X-ray scattering (SAXS) method is widely used in investigating protein structures in solution, but high-quality 3D model reconstructions are challenging. We present a new algorithm based on a deep learning method for model reconstruction from SAXS data. An auto-encoder for protein 3D models was trained to compress 3D shape information into vectors of a 200-dimensional latent space, and the vectors are optimized using genetic algorithms to build 3D models that are consistent with the scattering data. The program has been tested with experimental SAXS data, demonstrating the capacity and robustness of accurate model reconstruction. Furthermore, the model size information can be optimized using this algorithm, enhancing the automation in model reconstruction directly from SAXS data. The program was implemented using Python with the TensorFlow framework, with source code and webserver available from http://liulab.csrc.ac.cn/decodeSAXS.

Similarity Analysis of 3D Structures of Proteins Based Tile-CNN

The 3D structure of a protein is closely related to its function, and the similarity analysis between their structures can help reveal the function of proteins. However, there exist two problems arising from the analysis of 3D structures of proteins. The proteins with a similar sequence may have different structures, while the proteins with a similar structure may have different sequences. In the analysis of similarity in 3D structures of proteins, it remains difficult for the traditional methods using the spatial feature distribution and geometry or topology features of proteins to solve these problems. In this paper, a Tile-CNN network is proposed to analyze the similarity of proteins in 3D structure. In order to capture the overall and the local features as exhibited by the 3D structures of proteins, it projects 3D protein models into 2D protein images from different views and then cuts these 2D projected images using the tile strategy. After the training of proteins with these images in the Tile-CNN, the test protein model can be expressed by an analysis matrix, and then the similarity between 3D structures of proteins is computed using the root mean square distance (RMSD) for the benchmark matrix and the analysis matrix. As revealed by the experimental results, the proposed algorithm is more robust in analyzing the similarity of 3D structures of proteins and produces a satisfactory performance in solving the two aforementioned problems.

Comparative modeling and mutual docking of structurally uncharacterized heat shock protein 70 and heat shock factor-1 proteins in water buffalo

Aim:In this study, a wide range of in silico investigation of Bubalus bubalis (BB) heat shock protein 70 (HSP70) and heat shock factor-1 (HSF1) has been performed, ranging from sequence evaluation among species to homology modeling along with their docking studies to decipher the interacting residues of both molecules.Materials and Methods:Protein sequences of BB HSP70 and HSF1 were retrieved from NCBI database in FASTA format. Primary and secondary structure prediction were computed using Expasy ProtParam server and Phyre2 server, respectively. TMHMM server was used to identify the transmembrane regions in HSP70. Multiple sequence alignment and comparative analysis of the protein was carried out using MAFFT and visualization was created using ESPript 3.0. Phylogenetic analysis was accomplished by COBALT. Interactions of HSP70 with other proteins were studied using STRING database. Modeller 9.18, RaptorX, Swiss-Modeller, Phyre2, and I-TASSER were utilized to design the three-dimensional structure of these proteins followed by refinement; energy minimization was accomplished using ModRefiner and SPDBV program. Stereochemical quality along with the accuracy of the predicted models and their visualization was observed by PROCHECK program of PDBsum and UCSF Chimera, respectively. ClusPro 2.0 server was accessed for the docking of the receptor protein with the ligand.Results:The lower value of Grand Average of Hydropathy indicates the more hydrophilic nature of HSP70 protein. Value of the instability index (II) classified the protein as stable. No transmembrane region was reported for HSP70 by TMHMM server. Phylogenetic analysis based on multiple sequence alignments (MSAs) by COBALT indicated more evolutionarily closeness of Bos indicus (BI) with Bos taurus as compared to BI and BB. STRING database clearly indicates the HSF1 as one of the interacting molecules among 10 interacting partners with HSP 70. The best hit of 3D model of HSP70 protein and HSF1 was retrieved from I-TASSER and Phyre2, respectively. Interacting residues and type of bonding between both the molecules which were docked by ClusPro 2.0 were decoded by PIC server. Hydrophobic interactions, protein-protein main-chain-side-chain hydrogen bonds, and protein-protein side-chain-side-chain hydrogen bonds were delineated in this study.Conclusion:This is the first-ever study on in silico interaction of HSP70 and HSF1 proteins in BB. Several bioinformatics web tools were utilized to study secondary structure along with comparative modeling, physicochemical properties, and protein-protein interaction. The various interacting amino acid residues of both proteins have been indicated in this study.

Cellular Processes Involved in Jurkat Cells Exposed to Nanosecond Pulsed Electric Field.

Recently, nanosecond pulsed electric field (nsPEF) has been considered as a new tool for tumor therapy, but its molecular mechanism of function remains to be fully elucidated. Here, we explored the cellular processes of Jurkat cells exposed to nanosecond pulsed electric field. Differentially expressed genes (DEGs) were acquired from the GEO2R, followed by analysis with a series of bioinformatics tools. Subsequently, 3D protein models of hub genes were modeled by Modeller 9.21 and Rosetta 3.9. Then, a 100 ns molecular dynamics simulation for each hub protein was performed with GROMACS 2018.2. Finally, three kinds of nsPEF voltages (0.01, 0.05, and 0.5 mV/mm) were used to simulate the molecular dynamics of hub proteins for 100 ns. A total of 1769 DEGs and eight hub genes were obtained. Molecular dynamic analysis, including root mean square deviation (RMSD), root mean square fluctuation (RMSF), and the Rg, demonstrated that the 3D structure of hub proteins was built, and the structural characteristics of hub proteins under different nsPEFs were acquired. In conclusion, we explored the effect of nsPEF on Jurkat cell signaling pathway from the perspective of molecular informatics, which will be helpful in understanding the complex effects of nsPEF on acute T-cell leukemia Jurkat cells.

Ensembl 2020.

The Ensembl (https://www.ensembl.org) is a system for generating and distributing genome annotation such as genes, variation, regulation and comparative genomics across the vertebrate subphylum and key model organisms. The Ensembl annotation pipeline is capable of integrating experimental and reference data from multiple providers into a single integrated resource. Here, we present 94 newly annotated and re-annotated genomes, bringing the total number of genomes offered by Ensembl to 227. This represents the single largest expansion of the resource since its inception. We also detail our continued efforts to improve human annotation, developments in our epigenome analysis and display, a new tool for imputing causal genes from genome-wide association studies and visualisation of variation within a 3D protein model. Finally, we present information on our new website. Both software and data are made available without restriction via our website, online tools platform and programmatic interfaces (available under an Apache 2.0 license) and data updates made available four times a year.

In-silico molecular characterization and mutational analysis of inter-alpha-trypsin inhibitor heavy chain 4 in rheumatoid arthritis

Inter-alpha-trypsin inhibitor heavy chain (ITIH4) is IL-6-dependent acute phase protein. Increased expression of ITIH4 during acute phase inflammation of Rheumatoid Arthritis (RA) was reported earlier. Analyzing the heavy chain would, therefore, be crucial to understand the regulation and localisation of Serum-derived Hyaluronic acid-Associated Protein (SHAP) complex in acute phase inflamed conditions. In our previous studies, we have computationally constructed the 3D model of ITIH4 and checked its relevance. But further detailed structural annotation of ITIH4 is required to reveal its surface properties, domain characterisation, and molecular level interaction including drug designing. In the present study, we have utilized our constructed high-resolution three-dimensional (3D) model of ITIH4 protein to reveal the associated surface parameters and its functional role based on the structure-based homology search. Signal peptide prediction has been carried out by the Signalp 4.1 server, whereas COFACTOR server has been utilized for the prediction of active site residues, ligand binding residues, and gene ontology and enzyme classification. Immune epitope database (IEDB) analysis resource was used for the identification of antigenic peptides and Disease-Susceptibility-based SAV Phenotype Prediction (SUSPECT) server revealed the mutation prone residues of ITIH4. Present study indicated that mutation L304H and L304P may enable the protein model highly unstable. The molecular dynamic simulation was carried out for a comparative analysis between native and mutant protein. Additionally, interaction studies implicate that ITIH4 may modulate T cell-mediated inflammation in synovial region during RA. In-silico construction of ITIH4 mutants with single amino acid change may prove to be a tool in experimental studies to understand the importance of each functional domain.

Prevalence of ABCA4 Deep-Intronic Variants and Related Phenotype in An Unsolved "One-Hit" Cohort with Stargardt Disease.

We investigated the prevalence of reported deep-intronic variants in a French cohort of 70 patients with Stargardt disease harboring a monoallelic pathogenic variant on the exonic regions of ABCA4. Direct Sanger sequencing of selected intronic regions of ABCA4 was conducted. Complete phenotypic analysis and correlation with the genotype was performed in case a known intronic pathogenic variant was identified. All other variants found on the analyzed sequences were queried for minor allele frequency and possible pathogenicity by in silico predictions. The second mutated allele was found in 14 (20%) subjects. The three known deep-intronic variants found were c.5196+1137G>A in intron 36 (6 subjects), c.4539+2064C>T in intron 30 (4 subjects) and c.4253+43G>A in intron 28 (4 subjects). Even though the phenotype depends on the compound effect of the biallelic variants, a genotype-phenotype correlation suggests that the c.5196+1137G>A was mostly associated with a mild phenotype and the c.4539+2064C>T with a more severe one. A variable effect was instead associated with the variant c.4253+43G>A. In addition, two novel variants, c.768+508A>G and c.859-245_859-243delinsTGA never associated with Stargardt disease before, were identified and a possible splice defect was predicted in silico. Our study calls for a larger cohort analysis including targeted locus sequencing and 3D protein modeling to better understand phenotype-genotype correlations associated with deep-intronic changes and patients’ selection for clinical trials.

Molecular modelling and dynamics of CA2 missense mutations causative to carbonic anhydrase 2 deficiency syndrome

Carbonic anhydrase 2 (CA2) enzyme deficiency caused by CA2 gene mutations is an inherited disorder characterized by symptoms like osteopetrosis, renal tubular acidosis, and cerebral calcification. This study has collected the CA2 deficiency causal missense mutations and assessed their pathogenicity using diverse computational programs. The 3D protein models for all missense mutations were built, and analyzed for structural divergence, protein stability, and molecular dynamics properties. We found M-CAP as the most sensitive prediction method to measure the deleterious potential of CA2 missense mutations. Free energy dynamics of tertiary structure models of CA2 mutants with DUET, mCSM, and SDM based consensus methods predicted only 50% of the variants as destabilizing. Superimposition of native and mutant CA2 models revealed the minor structural fluctuations at the amino acid residue level but not at the whole protein structure level. Near native molecular dynamic simulation analysis indicated that CA2 causative missense variants result in residue level fluctuation pattern in the protein structure. This study expands the understanding of genotype-protein phenotype correlations underlying CA2 variant pathogenicity and presents a potential avenue for modifying the CA2 deficiency by targeting biophysical structural features of CA2 protein. Communicated by Ramaswamy H. Sarma

An immunoinformatics approach for design and validation of multi-subunit vaccine against Cryptosporidium parvum.

An immunoinformatics-based approach is explored for potential multi-subunit vaccine candidates against Cryptosporidium parvum. We performed protein structure based systematic methodology for the development of a proficient multi-subunit vaccine candidate against C. parvum based on their probability of antigenicity, allergenicity and transmembrane helices as the screening criteria. The best-screened epitopes like B-cell epitopes (BCL), Helper T-lymphocytes (HTL) and cytotoxic T- lymphocytes (CTL) were joined by using the appropriate linkers to intensify and develop the presentation and processing of the antigenic molecules. Modeller software was used to generate the best 3D model of the subunit protein. RAMPAGE and other web servers were employed for the validation of the modeled protein. Furthermore, the predicted modeled structure was docked with the two known receptors like TLR2 and TLR4 through ClusPro web server. Based on the docking score, the multi-subunit vaccine docked with TLR2 was subjected to energy minimization by molecular dynamics (MD) simulation to examine their stability within a solvent system. From the simulation study, we found that the residue Glu-107 of subunit vaccine formed a hydrogen bond interaction with Arg-299 of the TLR2 receptor throughout the time frame of the MD simulation. The overall results showed that the multi-subunit vaccine could be an efficient vaccine candidate against C. parvum.

Reconstruction of low-resolution molecular structures from simulated atomic force microscopy images

BackgroundAtomic Force Microscopy (AFM) is an experimental technique to study structure-function relationship of biomolecules. AFM provides images of biomolecules at nanometer resolution. High-speed AFM experiments produce a series of images following dynamics of biomolecules. To further understand biomolecular functions, information on three-dimensional (3D) structures is beneficial. MethodWe aim to recover 3D information from an AFM image by computational modeling. The AFM image includes only low-resolution representation of a molecule; therefore we represent the structures by a coarse grained model (Gaussian mixture model). Using Monte-Carlo sampling, candidate models are generated to increase similarity between AFM images simulated from the models and target AFM image. ResultsThe algorithm was tested on two proteins to model their conformational transitions. Using a simulated AFM image as reference, the algorithm can produce a low-resolution 3D model of the target molecule. Effect of molecular orientations captured in AFM images on the 3D modeling performance was also examined and it is shown that similar accuracy can be obtained for many orientations. ConclusionsThe proposed algorithm can generate 3D low-resolution protein models, from which conformational transitions observed in AFM images can be interpreted in more detail. General significanceHigh-speed AFM experiments allow us to directly observe biomolecules in action, which provides insights on biomolecular function through dynamics. However, as only partial structural information can be obtained from AFM data, this new AFM based hybrid modeling method would be useful to retrieve 3D information of the entire biomolecule.

Estimation of model accuracy in CASP13

Methods to reliably estimate the accuracy of 3D models of proteins are both a fundamental part of most protein folding pipelines and important for reliable identification of the best models when multiple pipelines are used. Here, we describe the progress made from CASP12 to CASP13 in the field of estimation of model accuracy (EMA) as seen from the progress of the most successful methods in CASP13. We show small but clear progress, that is, several methods perform better than the best methods from CASP12 when tested on CASP13 EMA targets. Some progress is driven by applying deep learning and residue-residue contacts to model accuracy prediction. We show that the best EMA methods select better models than the best servers in CASP13, but that there exists a great potential to improve this further. Also, according to the evaluation criteria based on local similarities, such as lDDT and CAD, it is now clear that single model accuracy methods perform relatively better than consensus-based methods.

Methods for the Refinement of Protein Structure 3D Models.

The refinement of predicted 3D protein models is crucial in bringing them closer towards experimental accuracy for further computational studies. Refinement approaches can be divided into two main stages: The sampling and scoring stages. Sampling strategies, such as the popular Molecular Dynamics (MD)-based protocols, aim to generate improved 3D models. However, generating 3D models that are closer to the native structure than the initial model remains challenging, as structural deviations from the native basin can be encountered due to force-field inaccuracies. Therefore, different restraint strategies have been applied in order to avoid deviations away from the native structure. For example, the accurate prediction of local errors and/or contacts in the initial models can be used to guide restraints. MD-based protocols, using physics-based force fields and smart restraints, have made significant progress towards a more consistent refinement of 3D models. The scoring stage, including energy functions and Model Quality Assessment Programs (MQAPs) are also used to discriminate near-native conformations from non-native conformations. Nevertheless, there are often very small differences among generated 3D models in refinement pipelines, which makes model discrimination and selection problematic. For this reason, the identification of the most native-like conformations remains a major challenge.

Using PconsC4 and PconsFold2 to Predict Protein Structure.

In spite of the fact that there has been a significant increase in the number of solved protein structures, structural information is missing for many proteins. Although structural information is codified in the amino acid sequence, computational prediction using only this information is still an unsolved problem. However, one successful method to model a protein's structure starting from the primary sequence is to use contact prediction derived from multiple sequence alignment (MSA). Here we use our contact predictor PconsC4 to generate a list of probable contacts between residues in the primary sequences. These contacts are then used together with the secondary structure prediction as constraints for the CONFOLD folding method. In this way, a 3D protein model can be built starting directly from the primary sequence. © 2019 by John Wiley & Sons, Inc.